General Language Research Articles

Visual Modeling Languages in Patient Pathways: Scoping Review.

Patient pathways (PPs) are presented as a panacea solution to enhance health system functions. It is a complex concept that needs to be described and communicated well. Modeling plays a crucial role in promoting communication, fostering a shared understanding, and streamlining processes. Only a few existing systematic reviews have focused on modeling methods and standardized modeling languages. There remains a gap in consolidated knowledge regarding the use of diverse visual modeling languages. This scoping review aimed to compile visual modeling languages used to represent PPs, including the justifications and the context in which a modeling language was adopted, adapted, combined, or developed. After initial experimentation with the keywords used to describe the concepts of PPs and visual modeling languages, we developed a search strategy that was further refined and customized to the major databases identified as topically relevant. In addition, we consulted gray literature and conducted hand searches of the referenced articles. Two reviewers independently screened the articles in 2 stages using preset inclusion criteria, and a third reviewer voted on the discordance. Data charting was done using an iteratively developed form in the Covidence software. Descriptive and thematic summaries were presented following rounds of discussion to produce the final report. Of 1838 articles retrieved after deduplication, 22 satisfied our inclusion criteria. Clinical pathway is the most used phrase to represent the PP concept, and most papers discussed the concept without providing their operational definition. We categorized the visual modeling languages into five categories: (1) general purpose-modeling language (GPML) adopted without major extension or modification, (2) GPML used with formal extension recommendations, (3) combination of 2 or more modeling languages, (4) a developed domain-specific modeling language (DSML), and (5) ontological modeling languages. The justifications for adopting, adapting, combining, and developing visual modeling languages varied accordingly and ranged from versatility, expressiveness, tool support, and extensibility of a language to domain needs, integration, and simplification. Various visual modeling languages were used in PP modeling, each with varying levels of abstraction and granularity. The categorization we made could aid in a better understanding of the complex combination of PP and modeling languages. Standardized GPMLs were used with or without any modifications. The rationale to propose any modification to GPMLs evolved as more evidence was presented following requirement analyses to support domain constructs. DSMLs are infrequently used due to their resource-intensive development, often initiated at a project level. The justifications provided and the context where DSMLs were created are paramount. Future studies should assess the merits and demerits of using a visual modeling language to facilitate PP communications among stakeholders and use evaluation frameworks to identify, modify, or develop them, depending on the scope and goal of the modeling need.

AltGosling: Automatic Generation of Text Descriptions for Accessible Genomics Data Visualization.

Biomedical visualizations are key to accessing biomedical knowledge and detecting new patterns in large datasets. Interactive visualizations are essential for biomedical data scientists and are omnipresent in data analysis software and data portals. Without appropriate descriptions, these visualizations are not accessible to all people with blindness and low vision, who often rely on screen reader accessibility technologies to access visual information on digital devices. Screen readers require descriptions to convey image content. However, many images lack informative descriptions due to unawareness and difficulty writing such descriptions. Describing complex and interactive visualizations, like genomics data visualizations, is even more challenging. Automatic generation of descriptions could be beneficial, yet current alt text generating models are limited to basic visualizations and cannot be used for genomics. We present AltGosling, an automated description generation tool focused on interactive data visualizations of genome-mapped data, created with the grammar-based genomics toolkit Gosling. The logic-based algorithm of AltGosling creates various descriptions including a tree-structured navigable panel. We co-designed AltGosling with a blind screen reader user (co-author). We show that AltGosling outperforms state-of-the-art large language models and common image-based neural networks for alt text generation of genomics data visualizations. As a first of its kind in genomic research, we lay the groundwork to increase accessibility in the field. The source code, examples, and interactive demo are accessible under the MIT License at https://github.com/gosling-lang/altgosling. The package is available at https://www.npmjs.com/package/altgosling. Supplementary data are available at Bioinformatics online.

Abstract 4135353: Accuracy, Hallucinations, and Misgeneralizations of Large Language Models in Reviewing Cardiology Literature

Introduction: The text generation abilities of large language models (LLMs), such as ChatGPT, can streamline the process of research and potentially expedite reviewing literature. Methods: We evaluated the capabilities of LLMs and their major susceptibilities—hallucination and goal misgeneralization—in reviewing cardiology literature. We asked 30 comprehensive questions (n=600) regarding findings from notable cardiovascular studies and randomized controlled trials (n=20) from 2016-2020 to GPT-4, GPT-3.5, and LLaMa 3. The question groups followed a format where the initial question checked for accuracy in identification, second for misgeneralization, and third for hallucination (Table 1). All questions were communicated in individual interfaces of the models to retain independence and prevent bias. Responses were then reviewed for accuracy, misgeneralizations, and hallucinations. Results: GPT-4 and GPT-3.5 did not significantly differ in accuracy (55% vs. 40%, p=0.527), with a low association (φ=0.1) and an odds ratio (OR) of 1.83 (95% CI: 0.522-6.434). Similar results were observed between GPT-4 and LLaMa 3 (55% vs. 30%, p=0.201), with an OR of 2.85 (95% CI: 0.776-10.467). For generalization, GPT-4 and GPT-3.5 also did not significantly differ (30% vs. 35%, p=0.999), with a low association (φ=0.001) and an OR of 0.80 (95% CI: 0.211-2.998). Likewise, GPT-4 and LLaMa 3 showed no significant difference (30% vs. 25%, p=0.999), with an OR of 1.29 (95% CI: 0.319-5.174). For non-hallucinations, GPT-4 and GPT-3.5 did not significantly differ (30% vs. 60%, p=0.187), with a low association (φ=0.214) and an OR of 0.33 (95% CI: 0.088-1.256). Lastly, GPT-4 and LLaMa 3 showed no significant difference (30% vs. 25%, p=0.836), with an OR of 1.50 (95% CI: 0.36-6.13). Overall, all models did not differ in accuracy, misgeneralization, and hallucination. Given the moderate susceptibilities of these models, LLMs require more training before realistic usage in cardiology. Conclusions: LLMs show potential for identifying cardiology studies, but moderate hallucination and misgeneralization render them unsuitable for this purpose. Without resolution of susceptibilities (<5-10%), LLMs remain unrealistic for accurately reviewing literature.

Large language models for generative information extraction: a survey

AbstractInformation Extraction (IE) aims to extract structural knowledge from plain natural language texts. Recently, generative Large Language Models (LLMs) have demonstrated remarkable capabilities in text understanding and generation. As a result, numerous works have been proposed to integrate LLMs for IE tasks based on a generative paradigm. To conduct a comprehensive systematic review and exploration of LLM efforts for IE tasks, in this study, we survey the most recent advancements in this field. We first present an extensive overview by categorizing these works in terms of various IE subtasks and techniques, and then we empirically analyze the most advanced methods and discover the emerging trend of IE tasks with LLMs. Based on a thorough review conducted, we identify several insights in technique and promising research directions that deserve further exploration in future studies. We maintain a public repository and consistently update related works and resources on GitHub (LLM4IE repository).

NCOG-32. RELATIONSHIPS BETWEEN INTRAOPERATIVE TASK-BASED GAMMA BAND ELECTROCORTICOGRAMS, DIRECT ELECTRICAL STIMULATION, AND COGNITIVE-LINGUISTIC OUTCOME FOLLOWING AWAKE BRAIN TUMOR RESECTION

Abstract BACKGROUND As an adjunct to direct electrical stimulation (DES), intraoperative electrocorticography (ECoG) can aid detection of language-eloquent cortex. However, it is unclear how surgical impact upon mapping findings relates to cognitive-linguistic outcomes. METHODS Ten patients underwent awake craniotomy for tumor resection with ECoG and DES language mapping. Mapping utilized iPad stimulus presentation coupled to an electrode array capable of both recording ECoG activity and delivering DES. Time-frequency gamma band activations were visualized (Early, Long, Late, Double Peak) for 3 naming paradigms (Object, Action, Auditory). DES was also conducted for each task. Mapping results were overlayed on postoperative MRI to examine proximity to parenchymal trauma (cavity, sulcal displacement). Patients completed cognitive-linguistic testing [WAB-R (N=3) or Neuropsychological Assessment (N=7)] prior to surgery and <2 months postoperatively. Patients with persisting aphasia at 6-month follow-up were also identified. RESULTS Postoperative language decline was found in 50% of patients, with 30% showing worsening (>1SD) in Naming and Phonemic Fluency, and 60% in Semantic Fluency. Surgery impacted ECoG features in 60% of patients and DES arrest sites in 30%. A trend was observed where Early and Late Object naming ECoG features were associated with Semantic Fluency decline [rpb(5)=-.69, p=.067]. Arrest during Auditory naming DES <1 cm of surgical effect was associated with decline in Abstraction [rpb(5)=-.83, p=.042; FET, p=.029] and Semantic Fluency [rpb(5)=-.87, p=.012; FET, p=.024]. Excluding the 3 transsulcal surgeries, arrest during Auditory naming DES <1 cm of resection was associated with general language decline [FET, p=.047]. All 3 patients with persisting aphasia had Auditory naming arrest <1 cm of resection and 2/3 patients had Object naming ECoG features resected. CONCLUSIONS While preliminary, results suggest that surgical proximity to ECoG and DES mapping results may inform risk of surgically-acquired cognitive-linguistic decline. Findings support the potential of multimethod-multitask intraoperative language mapping.

Smart Caption: Intelligent Image Description Using Transformer Decoder

The need for automated image description systems has grown significantly with the rise of multimedia content across diverse domains such as social media, digital libraries, and accessibility technologies. Smart Caption presents a novel approach for generating intelligent and context-aware image descriptions by utilizing a combination of Vision Transformers (ViTs) and Natural Language Processing (NLP) Transformer decoders. Unlike traditional convolution-based methods, Vision Transformers treat images as sequences of patches, enabling them to capture global image features more effectively. These extracted visual features are then processed by a Transformer-based decoder to produce coherent and contextually appropriate captions, bridging the gap between image recognition and natural language generation. Our approach focuses on leveraging the attention mechanisms inherent in both Vision and NLP Transformers to enhance the quality of image descriptions. The ViT architecture is designed to focus on relevant regions within an image, while the NLP Transformer decoder is adept at generating fluent and detailed descriptions by attending to the most significant visual features. This two-stage process improves the precision of object detection, relationship understanding, and scene context in generated captions. Experimental results demonstrate that Smart Caption out performs traditional methods in terms of accuracy, contextual relevance, and fluency, marking a significant step forward in the field of automated image captioning. Through this research, we aim to provide a scalable, efficient, and intelligent solution for image description, with potential applications in areas such as accessibility tools for visually impaired users, digital content indexing, and automated content generation. Our findings highlight the strengths of transformer-based models in bridging vision and language tasks, offering promising directions for future research in multimodal AI..

MeDi‐TODER: Medical Domain‐Incremental Task‐Oriented Dialogue Generator Using Experience Replay

ABSTRACTArtificial intelligence (AI) technology has brought groundbreaking changes to the healthcare domain. Specifically, the integration of a medical dialogue system (MDS) has facilitated interactions with patients, identifying meaningful information such as symptoms and medications from their dialogue history to generate appropriate responses. However, shortcomings arise when MDS lacks access to the patient's cumulative history or prior domain knowledge, resulting in the generation of inaccurate responses. To address this challenge, we propose a medical domain‐incremental task‐oriented dialogue generator using experience replay (MeDi‐TODER) that applies the continual learning technique to the medical task‐oriented dialogue generator. By strategically sampling and storing exemplars from previous domains and rehearsing it as it learns, the model effectively retains knowledge and can respond to the novel domains. Extensive experiments demonstrated that MeDi‐TODER significantly outperforms other models that lack continual learning in both natural language generation and natural language understanding. Notably, our proposed method achieves the highest scores, surpassing the upper‐class benchmarks.

Research on adversarial attack and defense of large language models

Abstract. Large language models (LLMs) have made excellent progress in text and image understanding and generation. However, with the wide range of applications of these models in various industries, the issue of their security, especially the defense against adversarial attacks, has become a focus of research. This study focuses on exploring the adversarial attacks faced by LLMs and their defense strategies, especially the design and optimization of defense mechanisms. Through literature review and case studies, this paper analyzes in detail the white-box and black-box attack patterns against LLMs, including model inversion, backdoor attacks, and token-based strategies. In response to these attacks, this paper proposes a series of defense strategies, including preventive measures such as data augmentation, adversarial training and model regularization, as well as real-time attack detection and response strategies such as anomaly detection and adversarial sample detection techniques. The core of this research is to improve the robustness and trustworthiness of LLMs, providing the necessary guarantees for their integration and sustainability in multiple industrial applications. In addition, this paper proposes future research directions, highlighting the importance of developing advanced defense systems, promoting interdisciplinary research and exploring new applications for LLMs. This research provides valuable insights into understanding and improving the security defense mechanisms of LLMs, which is essential for maintaining the security and user trust of these models.

Development and initial validation of the Cognitive Change Scale (CCS).

Cognitive impairment is common in neurologic diseases. Precise measurement of cognitive change over time is necessary for isolating disease-related patterns from normal age-related decline. Existing measures of subjective cognition, however, focus on present status. There is, to our knowledge, no currently available self-report measure of cognitive change. We therefore developed the Cognitive Change Scale (CCS), which assesses perceived cognitive change in neurologic populations. A systematic mixed-methods process was followed for the scale design and validation. Associations of CCS responses to demographics, mood, and fatigue were examined in 131 persons with multiple sclerosis. A total of 46 participants also completed a cognitive test battery. Correlations of test scores with CCS responses were calculated. The 17-item CCS showed good reliability and validity. Results of exploratory and confirmatory factor analyses supported a four-factor structure, with items reflecting change in (1) general cognition, (2) language and executive function, (3) external feedback, and (4) use of coping strategies. Positive relationships of CCS scores with fatigue, depression, and anxiety were observed. Correlations of CCS scores with cognitive test performance did not reach significance. The CCS may be a useful cognitive outcome tool for treatment trials in neurologic populations.

Primary school teachers’ views on language-responsive physical education

Language-responsive teaching aims to promote pupils’ participation opportunities by enabling them to acquire general academic language skills. It is assumed that the acquisition of academic language competencies must be accompanied and supported by the entire school and all school subjects. Yet, the idea of language-responsive Physical Education (PE) seems to be taken up only reluctantly in the current subject didactic and academic discourse. Thus, it is the aim of this article to contribute to bringing this topic into focus. The possibilities and limitations of language-responsive teaching in PE are examined from the teachers’ point of view. Special attention is paid to pupils facing language-barriers, as it is essential for this group to receive extra support to avoid educational disadvantages. Language education in PE has been a rather marginalized topic – much in contrast to the more recent discourse on general education and its crucial role for long-term academic and socio-economic success. It has been argued that the potential for language-building in PE has not been exploited satisfactorily. The observation that the idea of language-responsive PE has received little attention, despite its promising potential, begs the question of possible reasons. The following inquiry therefore investigates the perspectives of PE teachers on language-responsive teaching. The study consisted of 26 qualitative interviews with primary school teachers and aimed to generate insights into the importance attached to language teaching in PE in general (a), and its relevance for pupils’ participation in PE as well as for creating equal educational opportunities (b). Moreover, the study explores subject-specific potentials and challenges for language-responsive teaching faced by teachers with respect to pupils with special needs in language and communication or non-native speakers who have difficulties in understanding the language of instruction (c).

Physiological signal analysis and open science using the Julia language and associated software

In this mini review, we propose the use of the Julia programming language and its software as a strong candidate for reproducible, efficient, and sustainable physiological signal analysis. First, we highlight available software and Julia communities that provide top-of-the-class algorithms for all aspects of physiological signal processing despite the language’s relatively young age. Julia can significantly accelerate both research and software development due to its high-level interactive language and high-performance code generation. It is also particularly suited for open and reproducible science. Openness is supported and welcomed because the overwhelming majority of Julia software programs are open source and developed openly on public platforms, primarily through individual contributions. Such an environment increases the likelihood that an individual not (originally) associated with a software program would still be willing to contribute their code, further promoting code sharing and reuse. On the other hand, Julia’s exceptionally strong package manager and surrounding ecosystem make it easy to create self-contained, reproducible projects that can be instantly installed and run, irrespective of processor architecture or operating system.

«Языковая личность» - понятие и роль в современном мире

Relevance: In recent years, the Russian Federation has paid close attention to the official activities of employees of the internal affairs bodies from authorities, public organizations and institutions, and the population. First of all, we are talking about the observance by the employees of the internal affairs bodies of the norms of morality and ethics, rules of conduct in everyday life and speech culture. State policy determines the main areas of activity of the internal affairs bodies of the Russian Federation. Thus, in the decisions of the Ministry of Internal Affairs of the Russian Federation it has been repeatedly noted that the successful solution of the tasks facing the internal affairs bodies largely depends on the level of training of the personnel. Fostering high civic, professional, psychological and moral qualities in employees, as well as the formation of a high professional culture, one of the key competencies of which is proficiency in language as a means of communication, strengthening the authority of the internal affairs bodies in society, increasing the prestige of the service are priority goals in the activities of the police and are regulated by the order of the Ministry of Internal Affairs of Russia dated December 25, 2020 No. 900 "Issues of organizing moral and psychological support for the activities of the internal affairs bodies of the Russian Federation". Therefore, the problem of general language processes occurring in the internal affairs bodies of the Russian Federation at present leads to the need to study the features of communication underlying official activity, and therefore to its creator - the "linguistic personality". Statement of the problem: The article considers the issues of state support for traditional values ​​in the Russian Federation, the need to protect the norms of the Russian literary language, the role of language in human life. What are the features of the "linguistic personality"? One of the important conditions for the formation of a person's pedagogical culture is the study of pedagogical thought that has gone through a centuries-old historical path. The article considers the main and most significant stages in the development of the education system in the world, as well as domestic pedagogy, the founder of which is K.D. Ushinsky. Purpose of the study: clarification of the definition of "linguistic personality" based on the presented basic concepts of "linguistic personality" formed in world linguistics. Research methods: analysis of modern scientific definitions of the concepts of "personality", "linguistic personality", generalization of the results obtained. Results and key findings: the conducted research allowed us to formulate and propose the author's definition of the concept of "linguistic personality", which is proposed to be understood as a bearer of cultural and linguistic values, endowed with will and aspirations, moods and assessments, and also fully using language in business communication, professional activities and in everyday life.

Enhancing domain-specific text generation for power grid maintenance with P2FT.

The digitization of operation and maintenance in the intelligent power grid equipment relies on a diverse array of information for smart decision-making. In the domain of intelligent decision generation, proficiency is contingent upon extensive learning from copious amounts of text. This necessitates not only robust processing capabilities but also a high level of specialization. In addressing situations where authorization is lacking, pre-trained language models (PLMs) have already provided ideas when confronted with specialized domains or tasks. In consideration of the complexity of textual content in the field of the power grid, which encompasses a multitude of specialized knowledge and involves an abundance of proprietary terminology, we have undertaken an exploration of pre-trained model specialization using the power grid domain as an example, specifically for the task of generating maintenance strategies. A two-stage fine-tuning approach (P2FT) is employed, utilizing a large-scale pre-training model specifically designed for natural language processing. The efficacy and practical value of this method were evaluated through multiple metrics, juxtaposed with other advanced approaches involving low-parameter or parameter-free fine-tuning methods. Through a meticulous analysis and validation of experimental outcomes, we have corroborated the feasibility and practical application value of employing this approach for pre-trained model specialization. Additionally, it has furnished valuable guidance for text generation within both the Chinese language domain and the power grid domain.

KEY FEATURES OF PROFESSIONAL FOREIGN LANGUAGE COMMUNICATIVE COMPETENCE OF ENGINEERS

The authors analyze the approaches towards interpreting the notions of “foreign language competence”, “foreign language communicative competence”, and “professional foreign language communication”. The professional foreign language competence of the specialist is regarded as a result of the integration of professional and foreign language communicative competence. The interrelation between the university graduate's general communicative competence and professional foreign language competence is scrutinized. The components of foreign language communicative competence of engineering students are analyzed. The structure of this competence is based on the analysis of requirements set for the graduates by the job market. Cognitive and operational-technological competences comprising the professional section of the specialized part of the social-professional foreign language competence of the trainee engineer are relevant to the special competences of the linguistic section staying in a dynamic interaction with them. The offered model was successfully tried out during the interconnected learning of engineering subjects and a foreign language at Zaporizhzhia Polytechnic National University.

A Text Generation Method Based on a Multimodal Knowledge Graph for Fault Diagnosis of Consumer Electronics

As consumer electronics evolve towards greater intelligence, their automation and complexity also increase, making it difficult for users to diagnose faults when they occur. To address the problem where users, relying solely on their own knowledge, struggle to diagnose faults in consumer electronics promptly and accurately, we propose a multimodal knowledge graph-based text generation method. Our method begins by using deep learning models like the Residual Network (ResNet) and Bidirectional Encoder Representations from Transformers (BERT) to extract features from user-provided fault information, which can include images, text, audio, and even olfactory data. These multimodal features are then combined to form a comprehensive representation. The fused features are fed into a graph convolutional network (GCN) for fault inference, identifying potential fault nodes in the electronics. These fault nodes are subsequently fed into a pre-constructed knowledge graph to determine the final diagnosis. Finally, this information is processed through the Bias-term Fine-tuning (BitFit) enhanced Chinese Pre-trained Transformer (CPT) model, which generates the final fault diagnosis text for the user. The experimental results show that our proposed method achieves a 4.4% improvement over baseline methods, reaching a fault diagnosis accuracy of 98.4%. Our approach effectively leverages multimodal fault information, addressing the challenges users face in diagnosing faults through the integration of graph convolutional network and knowledge graph technologies.

Improving Factuality by Contrastive Decoding with Factual and Hallucination Prompts.

Large language models have demonstrated impressive capabilities in many domains. But they sometimes generate irrelevant or nonsensical text, or produce outputs that deviate from the provided input, an occurrence commonly referred to as hallucination. To mitigate this issue, we introduce a novel decoding method that incorporates both factual and hallucination prompts (DFHP). It applies contrastive decoding to highlight the disparity in output probabilities between factual prompts and hallucination prompts. Experiments on both multiple-choice and text generation tasks show that our approach significantly improves factual accuracy of large language models without additional training. On the TruthfulQA dataset, the DFHP method significantly improves factual accuracy of the LLaMA model, with an average improvement of 6.4% for the 7B, 13B, 30B, and 65B versions. Its high accuracy in factuality makes it an ideal choice for high reliability tasks like medical diagnosis and legal cases.

Bridging Human and Machine Intelligence: Reverse-Engineering Radiologist Intentions for Clinical Trust and Adoption

In the rapidly evolving landscape of medical imaging, the integration of artificial intelligence (AI) with clinical expertise offers unprecedented opportunities to enhance diagnostic precision and accuracy. Yet, the "black box" nature of AI models often limits their integration into clinical practice, where transparency and interpretability are important. This paper presents a novel system leveraging the Large Multimodal Model (LMM) to bridge the gap between AI predictions and the cognitive processes of radiologists. This system consists of two core modules, Temporally Grounded Intention Detection (TGID) and Region Extraction (RE). The TGID module predicts the radiologist's intentions by analyzing eye gaze fixation heatmap videos and corresponding radiology reports. Additionally, the RE module extracts regions of interest that align with these intentions, mirroring the radiologist’s diagnostic focus. This approach introduces a new task, radiologist intention detection, and is the first application of Dense Video Captioning (DVC) in the medical domain. By making AI systems more interpretable and aligned with radiologist’s cognitive processes, this proposed system aims to enhance trust, improve diagnostic accuracy, and support medical education. Additionally, it holds the potential for automated error correction, guiding junior radiologists, and fostering more effective training and feedback mechanisms. This work sets a precedent for future research in AI-driven healthcare, offering a pathway towards transparent, trustworthy, and human-centered AI systems. We evaluated this model using NLG(Natural Language Generation), time-related, and vision-based metrics, demonstrating superior performance in generating temporally grounded intentions on REFLACX and EGD-CXR datasets. This model also demonstrated strong predictive accuracy in overlap scores for medical abnormalities and effective region extraction with high IoU(Intersection over Union), especially in complex cases like cardiomegaly and edema. These results highlight the system's potential to enhance diagnostic accuracy and support continuous learning in radiology. We are also releasing the source code for our project, available here.

On the issue of the formation of the national identity of students (on the case of universities in Kazakhstan)

Introduction. The article examines the theoretical aspects and practical possibilities of formation of national identity in universities of Kazakhstan. This problem is becoming relevant in the context of the globalization of knowledge and international integration of education, the growth of academic mobility and the expansion of intercultural communications. Aim. The aim is to explore the mechanisms of constructing national identity and strengthening the national features of educational programs of higher education based on the study and generalization of the experience of Kazakhstani universities. Methodology and research methods. Based on the content analysis of normative documents on education, the content of bachelor's degree curricula and critical understanding of foreign scientific literature, the authors attempt to fill in the gaps in understanding the mechanisms of formation of national identity in higher education in the country and show its features. Results. The authors present the results of an analysis of the structure and content of bachelor educational programs at universities in Kazakhstan for the presence of the national component in them. Тhe authors summarize that the formation of national identity among the younger generation and filling it with semantic content is considered by the Kazakh academic community as an important state task, as a foundation for the cohesion and spiritual unity of citizens of a multinational country. The mechanisms of constructing national identity in the higher education system have been developed at the state level and are implemented in university training and educational practices. The national component of bachelor's degree programs, focused on formation of national identity, is provided by the content of general education humanities and language disciplines. University programs are developed in accordance with the requirements of the State Bachelor's Degree Standard, which contains a list of compulsory subjects for studying national history and culture, Kazakh, Russian and foreign languages. The principles of academic freedom allow universities to enhance the national flavor of educational programs by filling elective courses with national content. Scientific novelty. In modern conditions, the problem of ensuring the national identity of educational programs universities does not at all seem to be an axiom that contradicts the processes of internationalization of education. Universities around the world use various mechanisms to ensure the national identity of their programs, making them more attractive to students in the context of the development of intercultural communication. Kazakh universities are still gaining experience, but it may already be of interest from the point of view of accumulating practical experience. The practical significance of the study lies in the formulation of the problem as one of the important discourses in higher education. The results of the study will make a certain contribution to the understanding of the mechanisms and factors influencing the provision of national identity of educational programs using the example of a specific country. The article contributes to the scientific conversation about the meaning of national identity and can be a contribution to the study of the problems of forming the national identity of the younger generation.

A deep neural network model for classifying pharmacy practice publications into research domains

BackgroundPharmacy practice faculty research profiles extend beyond the clinical and social domains, which are core elements of pharmacy practice. But as highlighted by journal editors in the Granada Statements, there is no consensus on these terms. Four domains (clinical, education, social & administrative, and basic & translational) of pharmacy practice faculty research are proposed. ObjectivesTo develop a classifier for categorizing pharmacy practice faculty publications into four proposed domains, and to compare the model with zero-shot performances of state-of-the-art, general purpose large language models (gpLLMs). MethodsOne thousand abstracts from 2018 to 2021 documents published by pharmacy practice faculty were reviewed, labelled and used to screen and finetune several Bidirectional Encoders Representations from Transformers (BERT) models. The selected model was compared with zero-shot performances of 7 state-of-the-art gpLLMs including ChatGPT-4o, Gemini-1.5-Pro, Claude-3.5, LLAMA-3.1 and Mistral Large, using 80 randomly selected abstracts from 2023 publications labelled with ≥80% consensus by all authors. Classification metrics included F1, recall, precision and accuracy, and reproducibility was measured with Cohen's kappa. A use case was demonstrated by testing the null hypothesis that the research domain distribution of faculty publications was independent of the pandemic. ResultThe model – Pharmacy Practice Research Domain Classifier (PPRDC) produced a 5-fold stratified cross-validation metrics of 89.4 ± 1.7, 90.2 ± 2.2, 89.0 ± 1.7, and 95.5 ± 0.6, for F1, recall, precision and accuracy, respectively. PPRDC produced perfectly reproducible classifications (Cohen's kappa = 1.0) and outperformed zero-shot performances of all gpLLMs. F1 scores were 96.2 ± 1.6, 92.7 ± 1.2, 85.8 ± 3.2, and 83.1 ± 9.8 for education, clinical, social, and translational domains, respectively. ConclusionsPPRDC (https://sadeosun-pprdc.streamlit.app) performed better than gpLLMs in this abstract classification task. Among several other impacts, PPRDC opens a new frontier in bibliometric studies; it will also advance the goals of the Grenada Statements by aiding authors and journal editors in journal selection and article prioritization decisions, respectively.

Enhancing scene text detectors with realistic text image synthesis using diffusion models

Scene text detection techniques have garnered significant attention due to their wide-ranging applications. However, existing methods have a high demand for training data, and obtaining accurate human annotations is labor-intensive and time-consuming. As a solution, researchers have widely adopted synthetic text images as a complementary resource to real text images during pre-training. Yet there is still room for synthetic datasets to enhance the performance of scene text detectors. We contend that one main limitation of existing generation methods is the insufficient integration of foreground text with the background. To alleviate this problem, we present the Diffusion Model based Text Generator (DiffText), a pipeline that utilizes the diffusion model to seamlessly blend foreground text regions with the background’s intrinsic features. Additionally, we propose two strategies to generate visually coherent text with fewer spelling errors. With fewer text instances, our produced text images consistently surpass other synthetic data in aiding text detectors. Extensive experiments on detecting horizontal, rotated, curved, and line-level texts demonstrate the effectiveness of DiffText in producing realistic text images. Code will be available at: https://github.com/99Franklin/DiffText.